Semantic Segmentation using Cityscapes Dataset Semantic Segmentation: Image Segmentation or Semantic Image Segmentation is a computer vision task in which we label specific regions of an image according to what's being shown. In the image segmentation task, the model partitions the image into different segments, each of them representing a different entity. Hence, segmentating image in different parts which shows a semantic partition of the image.
Applications: Image segmentation is useful in many fields from self-driving cars, satellites, medical imaging. The subtleties in medical imaging are quite complex and sometimes even challenging for trained physicians. A machine that can understand these nuances and can identify the necessary areas can make a profound impact in medical care.
Dataset: For our project, we will be using cityscapes dataset which consists of diverse urban street scenes across 50 different cities at varying times of the year with dense pixel level annotations that allow us to build a semantic segmentation model. We have 5000 images at 1024x2048 resolution pre-split into train, validation, and test set and labeled annotations that span across 30+ classes and 8 categories that are commonly encountered during driving scene perception.
Youtube Video Link https://youtu.be/7yKbhTZifOw